By using perturbed projection matrix, a new descent memory gradient perturbed projection method for nonlinear Constrained optimization problems was developed. The global convergence properties of the new method were discussed. Combining FR, PR, HS parameters and the quasi-Newton equation with the new method, four new classes of memory gradient perturbed projection methods were presented. The classical conjugate gradient method is generalized for constrained optimization problems. The numerical results show that the new method is effective.%利用摄动投影矩阵建立求解非线性约束优化问题的记忆梯度摄动投影下降算法,并证明算法的收敛性,同时给出结合FR、PR、HS参数和拟牛顿方程的记忆梯度摄动投影算法,从而将经典的共轭梯度法推广用于求解约束优化问题.数值结果表明算法是有效的.
展开▼